Post-Cardiac-Arrest Coma: Multimodality Prognostication at 72 Hours
A 58-year-old man remains comatose 72 hours after witnessed ventricular fibrillation arrest with 22 minutes of resuscitation. He is intubated, sedated with propofol until 12 hours ago, and normothermic (36.6 C). Pupils are 3 mm and unreactive to bright light; corneal reflexes are absent; no cough with suctioning. Somatosensory evoked potentials after median nerve stimulation show absent N20 responses bilaterally on two separate runs with reproducible cervical potentials (N13 present). Serum neuron-specific enolase drawn at 48 and 72 hours is 38 and 44 mcg/L. EEG shows a continuous, unreactive, generalized suppression pattern at 3-5 microvolts with no sleep spindles or reactivity to stimulation. The family asks whether recovery is possible and whether a tracheostomy/PEG should be planned.
1. How do the SSEP findings, EEG, and neuron-specific enolase each contribute to prognosis, and what is the appropriate synthesis of these results?
Read the worked answer
SSEP: Bilaterally absent N20 responses with preserved cervical (N13) potentials indicate cortical/diencephalic injury with intact peripheral and lower cervical conduction. In post-arrest coma, absent N20 at 72 hours is one of the most robust single predictors of poor outcome, with a very low false-positive rate when the study is technically adequate (reproducible, no confound from limb ischemia, sedation, or hypothermia).
EEG: A continuous, unreactive suppression pattern without spindles or reactivity is a highly malignant background. It is not a burst-suppression pattern (which can be seen with sedatives), and the lack of reactivity on stimulation adds independent adverse weight.
NSE: Values of 38-44 mcg/L are markedly elevated (commonly cited threshold ~33 mcg/L), consistent with extensive neuronal injury; the rising trend strengthens the signal. NSE is less specific than SSEP (hemolysis can falsely elevate it), so it is supportive, not decisive, alone.
Synthesis: Multiple independent modalities (clinical brainstem reflexes, SSEP, EEG background/reactivity, NSE) converge on a very poor prognosis. The exam-relevant judgment is that concordant multimodal evidence, not any single test, justifies prognostication; with all four abnormal, expectation of meaningful neurologic recovery is very low and goals-of-care discussions (including avoidance of tracheostomy/PEG) are appropriate. Note the caveats: confirm normothermia, exclude residual sedation, and document technical adequacy of SSEP.
2. The family asks about a follow-up EEG and a repeat clinical exam at 96 hours. What is the value and limitation of serial testing here?
Read the worked answer
Serial testing can be useful to confirm that findings are stable rather than transient (e.g., resolving sedation or post-ictal suppression), and to detect delayed deterioration or subclinical seizures. However, once multimodal testing at 72 hours is concordant and severely abnormal, serial testing rarely changes the prognostic conclusion and should not be used to give false hope or to delay goals-of-care discussions indefinitely. The judgment point: serial exams are for confirming trajectory and detecting treatable complications (seizures, sedation effects), not for 'waiting out' a prognosis that is already established by concordant modalities.